Implicit conversion from float -11.731601731601728 to int loses precision